Our paper on Structural Pruning of Pre-trained Language Models via Neural Architecture Search got accepted at TMLR.